agentleFS
waynesuttonwaynesutton/builder-skills/skills/convex-security-check/SKILL.md

convex-security-check

Skill404 starsChanged 46 days agoApache-2.0View on GitHub

Ten minute security pass over a Convex backend: public functions that should be internal, missing auth checks, unvalidated args, IDs from the client trusted without ownership checks, secrets in code. Use before merging a pull request, after adding new public functions, or when the user says 'quick security check'.

  • Reads credentials

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run each grep, read what it surfaces, fix or file.

    line 10
  2. Very specific:

    Before merging a pull request that touches convex/

    line 14
  3. Distinctive:

    After adding or renaming public functions

    line 15
  4. Distinctive:

    Before a dev to prod push, when there is no time for a full audit

    line 17
  5. Very specific:

    Run the commands from the project root.

    line 21
  6. Distinctive:

    Stop and run the full audit when any of these are true:

    line 161

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Convex security check

A ten minute pass over convex/ that catches the mistakes that ship most often. The one rule: every exported query, mutation, and action is a public endpoint anyone can call with any arguments, so each one must check identity and ownership or be intentionally anonymous.

Run each grep, read what it surfaces, fix or file. This skill finds problems; it does not map the whole system. For that, hand off to convex-security-audit.

When to reach for this

  • Before merging a pull request that touches convex/
  • After adding or renaming public functions
  • The user says "quick security check", "sanity check the backend", or "anything obviously wrong here"
  • Before a dev to prod push, when there is no time for a full audit

Checklist

Run the commands from the project root. Every grep excludes _generated.

1. Public vs internal

Anything only called by other Convex functions, the scheduler, crons, or webhooks should be internalQuery, internalMutation, or internalAction.

rg -n "export const \w+ = (query|mutation|action)\(" convex --glob '!**/_generated/**'
rg -n "\bapi\.\w+\.\w+" convex --glob '!**/_generated/**'

The second grep finds api.* used inside the backend. Server code scheduling or running a public function is almost always a sign that function should be internal.

  • Every public function has a client that needs to call it
  • No api.* inside ctx.scheduler.*, ctx.run*, or crons.*
  • Admin only operations (role changes, credit grants, deletions across users) are internal

2. Auth

rg -l "= (query|mutation|action)\(" convex --glob '!**/_generated/**' \
  | xargs rg -L "getUserIdentity|getCurrentUser|authedQuery|authedMutation"

Files listed contain public functions and no auth call at all. Open each one. A public post list is fine anonymous. Anything that returns or writes user data is not.

  • Every public function checks identity or is documented as intentionally anonymous
  • Auth helpers throw or return early when identity is missing
  • Roles come from the users table, never from client arguments

3. Validation

rg -nU "(query|mutation|action)\(\{\s*handler" convex --glob '!**/_generated/**'
rg -n "v\.any\(\)" convex --glob '!**/_generated/**'

The first grep finds functions whose definition starts with handler, meaning no args. The second finds v.any(), which turns off validation for that field.

  • Every function has args and returns validators
  • No v.any() on arguments that reach the database or an external API
  • Return validators list fields explicitly so passwordHash, stripeCustomerId, and internal flags cannot leak

4. Ownership and IDs

v.id("tasks") proves the string is a valid ID for that table. It does not prove the caller owns the document.

rg -n "ctx\.db\.(get|patch|delete|replace)\(args\." convex --glob '!**/_generated/**'
rg -n "userId: v\.(id|string)\(" convex --glob '!**/_generated/**'

For each hit in the first grep, find the ownership comparison between the read and the write. For the second, a public function that accepts the caller's own userId as an argument is trusting the client to say who it is.

  • Every get, patch, delete, replace by a client supplied ID is followed by an ownership or membership check
  • Caller identity is derived from ctx.auth, never accepted as an argument
  • Lists use withIndex on the owner field, not a full scan plus filter

5. Secrets

rg -n -i "(sk_live|sk_test|whsec_|AKIA[0-9A-Z]{16}|-----BEGIN|api[_-]?key\s*[:=]\s*['\"][A-Za-z0-9])" convex src --glob '!**/_generated/**'
rg -n "process\.env\." src

Anything in src/ that reads process.env or import.meta.env ships to the browser. Only deployment URLs and public client IDs belong there.

  • No secret literals in convex/, src/, tests, or fixtures
  • Secrets read from process.env inside the action or HTTP action that uses them
  • Dev and prod deployments use different keys

6. HTTP and storage

rg -n "http\.route|httpAction\(" convex/http.ts
rg -n "storage\.(generateUploadUrl|getUrl)" convex --glob '!**/_generated/**'
  • Every http.ts route verifies its caller (webhook signature, bearer token, or getUserIdentity) before parsing the body
  • Routes call internal.*, not api.*
  • generateUploadUrl requires auth
  • getUrl is called on a storage ID read from a document the caller owns, never on a storage ID passed by the client

One fix, before and after

The most common finding: auth is checked, ownership is not.

Before:

export const updateTask = mutation({
  args: { taskId: v.id("tasks"), title: v.string() },
  returns: v.null(),
  handler: async (ctx, args) => {
    const identity = await ctx.auth.getUserIdentity();
    if (!identity) throw new ConvexError("Sign in required");
    // any signed in user can rename any task
    await ctx.db.patch(args.taskId, { title: args.title });
    return null;
  },
});

After:

import { mutation } from "./_generated/server";
import { v, ConvexError } from "convex/values";
import { getCurrentUser } from "./lib/auth";

export const updateTask = mutation({
  args: { taskId: v.id("tasks"), title: v.string() },
  returns: v.null(),
  handler: async (ctx, args) => {
    const user = await getCurrentUser(ctx); // throws when signed out
    const task = await ctx.db.get(args.taskId);
    // same error for missing and not owned, so IDs cannot be probed
    if (!task || task.userId !== user._id) {
      throw new ConvexError({ code: "NOT_FOUND", message: "Task not found" });
    }
    await ctx.db.patch(args.taskId, { title: args.title });
    return null;
  },
});

getCurrentUser looks the user up through a by_tokenIdentifier index using identity.tokenIdentifier and throws when there is no identity. Define it once in convex/lib/auth.ts and use it everywhere.

Common mistakes

MistakeWhy it breaksDo this instead
Checking auth in the React component onlyAnyone can call the function from the dashboard or a scriptCheck in the handler
v.id("users") argument for "the current user"Client can pass any user's IDDerive from ctx.auth.getUserIdentity()
Comparing ownership to identity.emailEmails can be reused or unverifiedCompare to user._id
internalMutation treated as safe on its ownThe public caller may pass unverified IDsCheck the call site too
Fixing one hit and moving onThe same pattern usually appears in siblingsFix all hits from the grep

Hand off to convex-security-audit

Stop and run the full audit when any of these are true:

  • More than two or three findings in steps 2 or 4; the pattern is systemic
  • The app has multi tenant data (orgs, teams, workspaces)
  • http.ts has webhook routes or routes that return user data
  • Files are uploaded and served
  • The user asks for a report, a launch review, or a post incident review

The audit maps auth per function, data access per table, HTTP exposure, storage, scheduler trust, rate limiting, and produces a written findings report. This check does not.

Docs

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.